The Calculation of Second Order Reaction Velocity Constants

Calculation of Second Order Reaction Velocity Constants. 699 in many sorption processes must involve activa- tion energies frequently of a considerabl...
0 downloads 0 Views 420KB Size
April, 1937

CALCULATION

OF

SECOND ORDER REACTION VELOCITY CONSTANTS

in many sorption processes must involve activation energies frequently of a considerable magnitude. It was a t first thought that the rate determining step was the approach of the gas molecule to the surface. That such could not universally be true was indicated by the work of Pace and Taylor* and by that of Kohl~chuetter~ demonstrating that, a t a pressure of one atmosphere and within limited temperature ranges, hydrogen and deuterium were adsorbed with the same velocity on zinc chromite, nickel, and chromium oxide gel, and by the work of Burwell and Taylor’O on the effect of pressure on the rate of the activated adsorption of hydrogen on chromium oxide gel. GarnerG*l1 has recently reiterated the contention that in general the slow process in activated adsorption is the “diffusion” of the gas down capillaries from an initial rapid adsorption, a view that derives initially from Ward.I2 The application of the term “diffusion” to the suggested mechanism is misleading since whereas true diffusion increases with the temperature as 1/T, the hypothesized (8) Pace and Taylor, J . Chem. Phys., 2, 578 (1934). (9) Kohlscbuettar, 2.physik. Chcm., AlTO, 300 (1934). 68, 697 (1936). (10) Burwell and Taylor, TKISJOURNAL, (11) Gamer and Veal, J . C h e m Soc., 1487 (1935). (12) Ward, Proc. Roy. Soc. (London), 8138, 522 (1931).

[CONTRIBUTION FROM TJX3

699

“diffusion” of Ward and Garner increases very much more rapidly. Lennard-Jones13 has shown that to account for the large temperature coefficient it is necessary to assume that before diffusion can take place, a molecule must be raised from its initially adsorbed state to a higher energy level. While the process proposed by LennardJones may occur in some instances, it cannot be a general one. For this hypothesis requires that the amount of adsorption be initially proportional to the square root of the time, a relation which rarely frts timeadsorption data. Further, such a picture is unable to account for the identity of the rates of adsorption of hydrogen and deuterium on the catalysts mentioned previously. Summary Experiments are reported which indicate that : (1) carbon monoxide is decomposed in the sense COZ, on surfaces of the equation, 2CO --+ C of zinc oxide a t temperatures above 184’. (2) Carbon deposited on the catalyst is removed by hydrogen a t temperatures of 184’ and above, presumably as methane.

+

(13) Lennard-Jones, Trans. Faraday Soc., 28, 333 (1932). PRINCETON, N.

J.

DEPARTMENT OF CHEMISTRY, Y A L E

RECEIVED OCTOBER 26, 1936

UNIVERSITY]

The Calculation of Second Order Reaction Velocity Constants BY JULIAN M STURTEVANT Introduction.-In the application of physical methods t o the determination of reaction velocities the observed quantity is usually a property of the reacting system which varies linearly with the concentrations of the reactants and prodpcts. To evaluate the two constants involved in this linear relation it is, in general, necessary to obtain the values of the observed property a t the start and finish of the reaction. In many cases it is impossible to obtain these values reliably; for example, if the reaction is being followed by observation of the temperature changes resulting from the heat evolved in the system under adiabatic conditions, there would obviously be serious experimental difficulties in obtaining the temperature of the system a t the completion of the reaction. A method of calculation which eliminates these difficulties in certain cases has been described by

Roseveare. This method involves the analytical or graphical elimination of the “infinity reading.” *4s pointed out by Roseveare, this method has the further advantage of not giving undue weight to the infinity reading in cases where it cannot be obtained with any greater accuracy than the other readings. Roseveare discussed the application of the method to first order reactions and to second order reactions in which the initial concentrations of the reactants are equal. In extending the method to more complicated reactions one finds that the algebraic treatment becomes excessively cumbersome. In the present communication it will be shown that these difficulties are avoided by a simple application of the methods of projective geometry. In the course of this treatment there will be developed a very easily applied test as to whether a given set of ( I ) Roseveare, THISJOURNAL, I S , 1651 (1931).

700

JULIAN

M. STURTEVANT

experimental data follows a certain kinetic equation. Kinetic Equation.-The general equations2 for several types of reactions may be written in the form a - x

kt = In c b - x

where k is proportional to the velocity constant, a # b, c are constants, and x is the extent of the reaction a t time t, or some quantity varying linearly with it. If x = xo a t t = 0, c = ( b - xo)/ (a - xo). Included among the types of reactions following this equation are bimolecular irreversible and reversible reactions (reverse reaction also bimolecular) in which the concentrations of the reactants are not equal, autocatalytic bimolecular irreversible reactions, and trimolecular irreversible reactions in which one reactant is a catalyst or its concentration is much greater than those of the other two reactants. In equation (1) the constants a, b, c will in general involve the initial concentrations of the reactants, the equilibrium constant for the reaction if it is reversible, and x~ and x , , the value of x at t = 03. Our problem is to determine k from the observed values of x without using xo, Xm*

Determination of k , a, b, c.-Proceeding in a manner analogous to that recommended by Roseveare‘ for first order reactions we suppose readings X I , x2, x3, x4 are taken at t, t At, t 2At, t 3At, readings x:, x i , x;, x i a t t’, t’ 2At, t’ 3At, and so on. Substitution At, t’ of these values in equation (1) and subtraction gives a series of simultaneous equations

+ +

+

+

+ +-

VOl. 59

/

The observationa that the quantity is the so-called anharmonic ratio4 or cross-ratio of the range of points a, b, X Z , xl suggests that the methods of projective geometry may find application here. The anharmonic ratio of a range of four collinear points, P I ,Pz,PQ, Pq may be defined as

In this notation equations ( 2 ) become [abxzxl]

[ab%&] = [abx4x8] = [abxixl] =

,

., =

a!

(5)

If we consider a, b as fixed points on a line the position of any other point Pi on the line is fixed by the ratio ai

= (a

It is easily shown that equations (5) become

- P i ) / ( b - Pi) [abPiPj]

(6) = ai/aj, so that

C Ya1= % a2 2 = crs 3 = a; E *L’ *

(7)

Using equations (6) and (7) and the definition of the anharmonic ratio we readily find that

. . . = 1 + a-1

[axlxzxs] = [axzxaxr] = [axix’xs’] =

[bxlxnxa] = [bxSxsx4] = [bxixix:] =

and -[xax1xax4] = -[xz’x1x;x4] =

*

.

= 1

+

(8)

(9)

... = a + 1 +

a-l

= constant

=A

(10)

Equation (10) gives us a simple method for calculating k and a criterion as to whether experimental data follow equation (1). Evidently

CY=-*d(T)* A - 1

A - 1

-1

(11)

- 1)/2)

(13)

.

2

where

so that a = constant (2)

from which a, b may be determined in terms of X I , xz, . . . Substitution of these values in any of the equations (3)

would enable us to obtain k. However, treatment of equations ( 2 ) by the usual algebraic methods leads immediately to cumbersome equations the simplification of which is difficult. (2) See the summary of equations, Vol. XI, in “Treatise O A Physical Chemistry,” edited by H. S. Taylor, D. Van Nostrand Company, New York, 1931, p. Q60.

KAt = In

CY

= cosh-l((A

According t o equation (12) data which conform to equation (1) should give a constant value of A > 3 . This test of the conformity of the data requires no knowledge of the concentrations of the reactants and, as stated before, x may be any quantity varying linearly with the concentrations. From the value of A the constant k may be calculated. The determination of the sign of the radical in equation (11) offers no difficulty, since (3) The author is indebted to Professor H. T. Engstrom of the Department of Mathematics for suggesting the use of projective geometry in this problem. (4) See, for example, “The Principles of Projective Geometry,” by J. L. S. Hatton, Cambridge University Press, 1913.

April, 1937

CALCULATION OF SECOND ORDERREACTION VELOCITY CONSTANTS

701

impossible t o obtain initial or final values of x. For example, in the determination of reaction velocities by following the rate of heat evolution6 it will evidently be out of the question to observe directly either the value of xo because of the heat of mixing, or of x, because of calorimetric difficulties. On the other hand, the concentrations of the reactants a t t = 0 will be known from the make-up of the reacting solution with an accuracy great enough to warrant their use in the manner k2At2 = ( A - 3) [l - ( A - 3)/12 . . . ] (14) described above. As a matter of fact, if it is posIn cases where a may be calculated accurately sible to make the original observations a t uniform by this method we may evaluate a, b from equa- time intervals, the present method of calculation may result in a considerable saving of time even tions (8) and (9) in cases where initial and final values of x can be axl(x8 - x2) - x8(xZ - x1) a = ’ obtained with good accuracy, both in enabling 4 x 3 - x:) - (x2 - x,) one to cease observations before the reaction is &(xS - x2) - a%(% - XI) b = (15) ( x 3 - x d - 4 x 2 - x1) finished and in simplifying the calculations. Example.-The data of Harned and Pfanstie16 and similar equations involving the other for the bimolecular reversible hydrolysis of ethyl Also c is readily obtained from equation (1) acetate catalyzed by acids have been treated by = a ( x s - x2) - (x2 - XI) -k W / A t (16) the above method. These data follow equation (x2 - xn) - 4 x 2 - x1) In such cases it would thus be possible in general to (I) very closely when calculated in the usual way. In this particular case of a bimolecular reversievaluate one of the initial concentrations or the ble reaction the constants k and c are given by equilibrium constant from the observed data, in addition to xo and x, . k = K d ( A - B)2 + 4 A B C (17) The magnification of errors involved in equation A B d ( A - B ) 2 4ABC c = (18) (11) in many cases may be avoided by using the A B - d ( A - B)’ 4ABC experimentally determined values for the initial concentrations (and the equilibrium constant if where A, B are the initial concentrations of water the reaction is reversible). In general it will be and ester, K is the velocity constant (natural possible to calculate one of the quantities a, b, logarithms) and C is the equilibrium constant. c (or some combination of them) accurately from Naturally K and c have different values with rethe known initial concentrations. Then one actions of other types. The quantity x in this merely has to calculate a from the appropriate case is the titer of the reacting solution, and may be one of equations (13) and (16) (or the appropriate observed in terms of alkali of unknown strength. combination of them). If c is the constant whose Since there is acid present as catalyst xo > 0. The experimental data’ given by Harned and value is considered as being accurately known, CY will have to be calculated by successive approxi- Pfanstiel for one run were plotted on a large scale, mations, the first of which may be obtained from and four sets of values of x were read off the plot, with At = 300 minutes. These sets gave the equation (11). In connection with this method of calculation values A = 3.153, 3.144, 3.142, 3.110. The trend two things may be pointed out. It is only neces- observed here is possibly somewhat outside the sary to have three readings XI, XZ,x3 in each set so experimental error, though it should be observed that the time interval At may be taken larger. that the quantity x4 - xs ( = 13) is the difference Also, in cases where a is very sensitive to the value of considerably larger quantities (= 60). At any of A , there will be a corresponding sensitivity of rate, it is obvious that a satisfactory value of K c and either a or b to a, so that the value of a could not be otained from these readings. The (5) This method of determination will be the subject of a future may be obtained accurately. communication. It may be emphasized again that this extension (6) Harned and Pfanstiel, THIS JOURNAL, 44, 2194 (1922). (7) The quantity recorded by Harned and Pfanstiel is x - X O , of Roseveare’s method of calculation will be of but since only differences in x appear in the calculations we may use greatest importance in cases where it is physically x - xo in place of x.

the quantity with the plus sign is the reciprocal of that with the minus sign, and one can readily tell whether KAt should be positive or negative. Unfortunately there will be a serious magnification of errors in cases where A is not much larger than 3, since a varies rapidly with A in this vicinity. Means of avoiding this difficulty are described below. In passing it may be pointed out that expansion of the inverse hyperbolic cosine gives the expression

+

XIS.

+ + +

+ +

702

G. P. BAXTER, J. H. FAULL, JR.,

mean value of A leads to K = 2.20 X loF5, while Harned and Pfanstiel found K = 2.342 X It is thus necessary to calculate K from the known value of c. The constants a, b involve X O , x , and therefore cannot be used. Four sets of three readings each were interpolated with At = 400 minutes. From these sets, together with the value c = -39.86 calculated from the initial concentrations and equilibrium constant, the values a = 1.701, 1.700, 1.698, 1.691 were calculated by successive approximations. The trend here, amounting to a total of 0.6%, may be partially due to inaccuracies introduced by the graphical interpolation. From the mean value of a! one obtains K = 2.365 X which is 1% higher than the value of Harned and Pfanstiel. This method of calculation eliminates the necessity of calculating the initial and final values of the titer as done by Harned and Pfan-

[CONTRIBUTIONFROM

THE

AND

F. D. TUEMMLER

VOl. 59

stiel. It should be emphasized that the value of a! would be changed by less than 0.1% by a change of 5% in c. S U V An application of the methods of projective geometry to a series of measurements of a property x (of a system reacting in accord with the kinetic equation Kt = In c (a - x ) / ( b . - 2)) which varies linearly with the progress of the reaction, taken at uniform time intervals, At, gives the result A

=Xa ZZ

- XZ

- x1

z=-z X4

XI

=

1

+ 2 cosh (k At)

By means of this result the necessity of obtaining initial and final values of x is eliminated. The method is illustrated using the data of Harned and Pfanstiel on the hydrolysis of ethyl acetate. NEW'HAVEN,

CONN.

RECEIVEDJANUARY 2, 1937

T. JEFFERSON COOLIDGE, JR., MEMORIALLABORATORY, HARVARD UNIVERSITY]

The Atomic Weights of Some Radiogenic Leads BY G . P. BAXTER,J. H; FAULL, JR.,'

In connection with the programs of the Committee on Geophysical Research of Harvard University and the Committee on Determination of Geologic Time of the National Research Council the atomic weights of some new specimens of radiogenic leads have been determined. Samarskite from Glastonbury, Connecticut, U. S. A.The original ground sample, weighing 1130 g., had been cleaned by gravity and magnetic methods by Mr. Harry Berman of the Harvard Mineralogical Laboratory. A preliminary analysis of a part of this material by R. C. Wells2 shows 6.91% uranium, 3.05% thorium, 0.311% lead. Treatment with redistilled hydrofluoric acid in platinum followed by repeated evaporation with nitric acid to remove the hydrofluoric acid and extraction of the insoluble residue with nitric acid, all in platinum, was carried out by Dr. C. M. Alter. The nitric acid solution on standing for several months deposited much columbic acid (40 g.). The dried insoluble residue weighed 970 g. This residue was found spectroscopically to contain most of the lead. Lead was separated from the nitric acid solution by double precipitation of basic acetates and treating the (1) Research Associate in Geology in Harvard University, 1935. The present work is Paper No. 33 published under the auspices of the Committee on Geophysical Research and of the Division of Geological Sciences at Harvard University. (2) R . C. Wells, Reports of the Committee on Determination of Geologic Time, of the National Research Council, April, 1935, and Sept., 1936.

AND

F. D. TUEMMLER

filtrates with hydrogen sulfide. The original insoluble residue and that which separated from the nitric acid solution were digested for long periods with redistilled hydrochloric acid and the lead precipitated from the extract after neutralization with ammonium acetate. About 0.1 g. of lead was obtained from the extract, 1.6 g. from the insoluble residues, together representing about half the lead content of the original mineral. Purification included the following processes : conversion of sulfide to sulfate through the chloride, solution of the sulfate in ammonium acetate and reprecipitation with hydrogen sulfide, solution of the sulfide in hydrochloric acid and conversion of the chloride to nitrate by repeated evaporation with nitric acid, four recrystallizations of the nitrate from nitric acid in quartz, conversion of nitrate to chloride with hydrochloric acid in quartz, four recrystallizations of chloride in quartz, and distillation of the chloride in dry hydrogen chloride in a quartz tube. All the product was used in one analysis. Lead was recovered from the solution of lead nitrate resulting from the first analysis and from the mother liquors remaining from the purification by precipitation as sulfide, and after conversion to the nitrate through the chloride it was re-purified as above. Beaverlodge Lake Pitchblende.-This specimen was collected and contributed by Dr. Fred Joliffe from a quartz vein near Beaverlodge Lake, Northwest Territories, Canada, about one hundred miles south of Great Bear Lake. The mineral consists chiefly of an intimate mixture of pitchblende and specularite. Weighted averages from analyses of five lots of similar material, representing 3000